Re: stability of replication?
On Thu, Sep 06, 2001 at 12:36:12PM -0700, Duc Chau wrote: I just wanted to know how stable is mysql databse replication is? What we have in mind is a tiered system. $master (california)-(internet ssh tunnel) \$slave (texas) 10 other slave machines on the same network (running 5 separate db per machine) \$slave (san fransisco) 10 other slave machines on the same network (running 5 separate db per machine) \$slave (new york) 10 other slave machines on the same network (running 5 separate db per machine) How fesible would this be? From an architectural standpoint, it will work. I'm not sure what you mean by running 5 separate db per machine, though. and has anyone worked with replication on a large scale? In terms of the number of slaves, or the distance between them? We have some West Coast (USA) -- East Cost replication happening, and it works rather well. The number of slaves won't be a problem. I've not run 10 at a time yet, but slaves really aren't much of a drain on the master. We're dealing with 2500 tables about 65k total records Read/Write to the master will be like 50 per second How many of those 50 are writes? Any input or links to documentation is much appreciated. Other than the replication docs in the manual, I can't suggest much else. But if you find stuff that needs to be added, speak up. :-) Jeremy -- Jeremy D. Zawodny, [EMAIL PROTECTED] Technical Yahoo - Yahoo Finance Desk: (408) 349-7878 Fax: (408) 349-5454 Cell: (408) 685-5936 MySQL 3.23.41-max: up 1 days, processed 17,193,429 queries (179/sec. avg) - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
stability of replication?
I just wanted to know how stable is mysql databse replication is? What we have in mind is a tiered system. $master (california)-(internet ssh tunnel) \$slave (texas) 10 other slave machines on the same network (running 5 separate db per machine) \$slave (san fransisco) 10 other slave machines on the same network (running 5 separate db per machine) \$slave (new york) 10 other slave machines on the same network (running 5 separate db per machine) How fesible would this be? and has anyone worked with replication on a large scale? We're dealing with 2500 tables about 65k total records Read/Write to the master will be like 50 per second Any input or links to documentation is much appreciated. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: stability of replication?
I have never tested with your magnitude - no where near that. You obviously would have one maching connected to the master per sub-network-cluster. [thus calling it tiered =]. I have tested over a network, using artificial packet destuction (for good measure), from a Linux box, running the master, to another Linux box running pptp, to a Win2k box (of course via pptp), with a slave server, then to one more win2k box as a slave. I had no trouble at all. The one caveat that I would mention is that I would occasionally have a bad sql statement, that was painful. I had to go from host to host doing SET SQL_SLAVE_SKIP_COUNTER=n; slave start [http://www.mysql.com/doc/R/e/Replication_SQL.html] That was painful. Also, using concurrent inserts is unsupported, i.e.: insert into table b select * from table A -or- create table dump as select important_data from table A == Alias locking is not quite the same either, and I use lots of that. But, on a whole, my new setup will use replication, things that formerly used concurrent selects will now be using a less elegant setup. I have servers in 2 levels: master slave1 slave2 (eventually I will have an online testing system that will plug in below slave2.) I use master for writes reporting queries, slave 1 for reads, and slave 2 for alias locks. In my pre-live testing scenario everything work pretty well. + It keeps everyone working well. hope all this swallop helps. js On Thursday 06 September 2001 01:36 pm, you wrote: I just wanted to know how stable is mysql databse replication is? What we have in mind is a tiered system. $master (california)-(internet ssh tunnel) \$slave (texas) 10 other slave machines on the same network (running 5 separate db per machine) \$slave (san fransisco) 10 other slave machines on the same network (running 5 separate db per machine) \$slave (new york) 10 other slave machines on the same network (running 5 separate db per machine) How fesible would this be? and has anyone worked with replication on a large scale? We're dealing with 2500 tables about 65k total records Read/Write to the master will be like 50 per second Any input or links to documentation is much appreciated. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php